|
In regards to not having skill causing strategy to leave a game is ridiculous. Arguably C. Ronaldo who plays for Manchester Untied has some of the most gifted and imaginative play in the game currently. Does this mean Manchester United implements the most strategy in a game? Arsenal and Barcelona use less strategy because they do not field C. Ronaldo?
Having improved skills may make certain strategies more viable, but not having those skills does not mean you cannot use strategy. I for example play zerg. My muta harassment does not mimic Jeadong.. it doesn't come close. In fact it fails so miserably that I instead use strategies that work with my skill set. I don't completely forsake strategy. I can't rationalize why not having skills necessary for some strategies means you completely use no strategy.
So please enlighten us to how we can find merit in the argument that not having a skill means you do not have any strategy so we may offer counter-arguments.
|
he's what we call a "scrub".
Starcraft just isn't for the wuss of heart.
|
On May 06 2008 17:53 InterWill wrote: StarCraft is a hard game. Further increasing the game speed just adds insult to injury for new players. This is what Dan is trying to get at. Saying that players aren't worthy, or that they should play another game, especially when a new game is under development is an odd reaction.
There are many levels of play in StarCraft, as well as many other competitive games. Most players never get past the lowest levels. These are things like playing against computer players or playing money maps. These levels are called the lowest because they have no real barrier of entry: anyone can play at them. They take no skill at all, you just mass units then attack. Some of these games don't even allow rushing! Above those levels are the private ladders that we have today. There is a bit of a skill barrier here, as you have to know your shit to get anywhere in these ladders. At the very top is pro gaming, where you train day and night with the best of the best, and your games are broadcast all around the globe. StarCraft is only a hard game if you play it above the lowest levels.
Now, back to Dan's article. I think Dan's article is basically the result of culture shock. He claims to have played StarCraft for a long time, which is true, but has he been playing it on a high level of play? I think that he, like many others before him, thinks that, logically, since he's played for a long time, he's really good. This is the mentality of a lot of gamers nowadays; I suspect it stems from games like WoW. But I digress. He seems to think that simply playing for a long time means he should be really good at the game, and prior to getting his shit stomped by LR, he had no experience playing against someone above his level of play, so this reinforced his belief. The fact that he got stomped by an attack force three times the size of his during that open SC2 match is proof that he, in comparison to LR, was on a lower level of play. However, instead of conceding that he was playing against someone who's slightly above his level of play, he claimed, in his editorial, that he was up against a professional, against someone that played SC for a living. He even went so far as to say that LR won only because LR clicked faster and he also implied that Dan could have won if Dan had "out-thinked" LR. However, this isn't the case. Dan didn't lose because of any of these reasons, he lost because LR was playing on a higher level of play, and it's a level that Dan and many others like him will never reach.
Dan is a busy guy. He makes his money from writing, not gaming, so I'm gonna guess that that he only plays any given game long enough to get a superficial grasp of the game and then spends twice the time he spent on said game on his review of that game. After all, he has to go on to the next game ASAP: he's got deadlines to meet. This is what all video game reviewers do nowadays, and it's one of the reasons why I don't trust professional reviewers anymore. Getting back to the point, because Dan never gets past the lowest levels of play, he's probably never met a person who has. Sure, he's heard the occasional anecdote of some god-like Korean gamer, and he's probably heard of the pro scene, but he's never experienced SC or any other RTS above its lowest levels of play because of time commitments.
So, when he was invited to play SC2 against LR, there was some culture shock. LR, I'll wager, has played a shit-ton more StarCraft than Dan ever will, and against opponents who know how to play the game. LR used this prior experience and applied it to SC2, resulting in Dan getting his shit stomped. After that, Dan, instead of asking, "Hey, how'd you do that?" made up a list of excuses how his opponent (in his mind) cheated and concluded that SC2 itself is flawed. This type of mentality is the mentality of what fighting gamers call a "scrub".
Sadly, Dan is probably never going to admit any of this because he knows that his reader base is a lot larger than the SC fandom, so he can therefore safely ignore us. He's going to continue being blind while leading the blind. He knows that he can just bitch about "clickfests" and can continue touting how he could totally win if he just "out-thinks you" because he's a scrub, his audience is scrubs and it's easier to blame someone else for a loss than it is to take responsibility.
|
On May 07 2008 01:40 yare wrote: So please enlighten us to how we can find merit in the argument that not having a skill means you do not have any strategy so we may offer counter-arguments. edit to "as many strategy's" and you yourself say it's true.
here's my two cents, I address only the question "Does increasing game speed reduce strategy?", not "Are fast games more fun?".
The speed the game is played at does change the strategies that can be effectively used, in someways it limits them, like not having good muta control means you can't win with a muta harass the way july would or having poor goon control means i can't take out spider mines the way good players can.
But it opens up more options as well, like if i know he's unlikely to spot the tiny shuttle dot on his mini map right away my storm drop is so much more effective, a flaw in even the highest level of progamers. Or you know that if you invest your time in micro you don't need to build as large an army (or if you micro for these 2 seconds you can leave your factories not producing), this is a strategy which is not in TBS games.
It introduces a new level of decision making, you have to decide which, out of the set of actions you should take, are more important. Where as in a slower game you only have to decide which are the actions you should take. These decisions are, to most people, less exciting.
Unlike speed chess where the quality of matches (compared to slow chess) is impaired by the players ability to analyze a known situation and think of a solution fast enough, the quality of play in fast SC compared to slow SC is inhibited by a lack of information (not noticing the red blobs on the mini map, not having time to count the number of zerglings or find the marine with less hp to focus fire on), hand speed and time spent decision making. Of these decision making is the least time consuming and decreasing the speed of the game even very drastically won't effect that.
If you say a better strategist is a player who, from the limited set of sensible options, picks better ones, then the speed of the game has very little effect on the winner in a game between two otherwise equally matched players. However, hand speed and reaction time have a massive effect on the strategies that are available and increasing the speed changes (and sometimes decreases) the set of viable strategies at any one time.
btw compare this thread to: Faster than fastest?
oh and his statement about min maxers is total rubbish, the faster the game the less perfect the play so the less accurate the model. Discard that from any further discussion, but the question "Does increasing game speed reduce strategy?" is interesting.
edit: I think exo6yte is totally right about Dan, low level player vs high level player, lost hard and it hurt, LR's strategy was way better than his as LR has played so much more SC.
|
Chess has no strategy. All people do is trade pieces until somebody hangs a piece, and then that player loses. Although this statement may be true for quite a lot of people, and somehow it might qualify as an "opinion," it doesn't stop me from being insanely, completely wrong.
|
drift: i agree if that were the statement. the statement in question is playing a game faster throws strategy out the window (in other words, NO strategy).
|
yare: well you have to expect Dan the exaggerate a little, being mildly outrageous makes the article more spicy if less accurate.
|
On May 07 2008 05:54 EmeraldSparks wrote: Chess has no strategy. All people do is trade pieces until somebody hangs a piece, and then that player loses. Although this statement may be true for quite a lot of people, and somehow it might qualify as an "opinion," it doesn't stop me from being insanely, completely wrong.
Agree...lol that was a good post.
The fact of the matter is that people have this annoying habit of tossing their ignorant opinions upon others. Personally, I get embarrassed if I talk out of my ass and am completely wrong about something - heck, I'll usually apologize for it. But most people aren't like this, so don't expect to see many good articles on SC2 from North America. Game reviewers are all casual gamers, and if their job is to relate to all the other casual gamers, then that's that.
|
After having watched GGPlay vs NobodyZerg, I think I've decided slower speeds would actually increase tactics, not strategy at all. When NobodyZerg attacked GGPlays base and they both had Mutalisk/Scourge masses, they would both have been able to split the scourge a lot better at a slower speed.
|
to anyone who has some high level undergraduate discrete mathematics courses in operations research:
excel can't exactly vomit out the optimal solution with a handful of numbers. you have to cleverly set up a linear program w/ the proper constraints. since linear programs can be solved in polynomial time, this is good news.
however, its extremely difficult to setup a linear program to solve an RTS problem. for example, resources are particularly hard to model. For example, the rate of mineral increase is a step function that behave in a linear fashion (when you have more mineral patches than workers) or in a logarithmically diminishing pattern (when you slowly have more SCVs than minerals) and have seemingly peculiar spikes/changes (when you transfer your workers to an expansion). As a result, even should you obtain an excellent model for resource collection, you still have a non-linear constraint on your hands, yielding a non-linear programming problem.
moreover, since RTS game are inherently discrete, you additionally have an integer programming problem. Occasionally you can circumvent this problem by letting your variables be time variables denoting the time at which unit/building X is constructed. Naturally, your objective function will be minimizing the time of the "last" built unit. However, you'll have a bit of an issue with "army" variability, as you'll need to specify exactly which units/buildings are going to be built (remember your variables are time variables).
I spent much of last semester constructing a linear programming problem to construct the optimal fast tank push build. My team accounted for the integer nature of the problem, and used the "time variables" i mentioned in the previous problem. That said, we had to assume infinite resources, as it was far too difficult (given our time span) to appropriately model a resource function. Even if we DID get a good model, we'd have a non-linear integer programming problem which is a real bitch to solve. Even with our simple model of "you have infinite resources, we need to build these specific units/buildings in the minimum amount of time," we still had alot of difficulty properly inputting our constraints.
I'd love to look at some of these "magical excel spreadsheets" that produce optimal values. But, considering the experience I have in modelling such things, I can't quite see how such a task would be feasible unless the player was a talented mathematician/engineer.
Thoughts?
|
On May 07 2008 09:50 Day[9] wrote: to anyone who has some high level undergraduate discrete mathematics courses in operations research:
excel can't exactly vomit out the optimal solution with a handful of numbers. you have to cleverly set up a linear program w/ the proper constraints. since linear programs can be solved in polynomial time, this is good news.
however, its extremely difficult to setup a linear program to solve an RTS problem. for example, resources are particularly hard to model. For example, the rate of mineral increase is a step function that behave in a linear fashion (when you have more mineral patches than workers) or in a logarithmically diminishing pattern (when you slowly have more SCVs than minerals) and have seemingly peculiar spikes/changes (when you transfer your workers to an expansion). As a result, even should you obtain an excellent model for resource collection, you still have a non-linear constraint on your hands, yielding a non-linear programming problem.
moreover, since RTS game are inherently discrete, you additionally have an integer programming problem. Occasionally you can circumvent this problem by letting your variables be time variables denoting the time at which unit/building X is constructed. Naturally, your objective function will be minimizing the time of the "last" built unit. However, you'll have a bit of an issue with "army" variability, as you'll need to specify exactly which units/buildings are going to be built (remember your variables are time variables).
I spent much of last semester constructing a linear programming problem to construct the optimal fast tank push build. My team accounted for the integer nature of the problem, and used the "time variables" i mentioned in the previous problem. That said, we had to assume infinite resources, as it was far too difficult (given our time span) to appropriately model a resource function. Even if we DID get a good model, we'd have a non-linear integer programming problem which is a real bitch to solve. Even with our simple model of "you have infinite resources, we need to build these specific units/buildings in the minimum amount of time," we still had alot of difficulty properly inputting our constraints.
I'd love to look at some of these "magical excel spreadsheets" that produce optimal values. But, considering the experience I have in modelling such things, I can't quite see how such a task would be feasible unless the player was a talented mathematician/engineer.
Thoughts? You do realize that Dan just made that whole excel-sheet story up right?
Even using the term "Min/Maxer" is pretty stupid, because you might as well replace it with "good player" or "pro".
|
First time poster, long time reader. Just wanted to respond to the thread with a few of my thoughts.
Point 1) First, on the issue of strategic depth vs speed. I'm pretty sure that I won't need to define the word 'strategy' to anyone within this thread, but what does this strategy actually consist of in-game? Please, something we can count; and that would be the actions performed. We can make the case that the actions a player takes within a game correspond with his strategy; furthermore, we can make the corollary that for each 'strategy' there is a specific series or pattern of actions that takes place. You can obviously expand this as you like, but for my purposes, all we need to remember is that there are some number of strategies that are correlated by a specific series of actions. Let's call the number of strategies X.
Begin Text for Clarification
Alright, so, a lot of you are defining 'strategy' as 'general overall flow of the game.' However, we can equate this to what actions we perform (we take with it the assumption that all actions taken in a game of StarCraft correspond to the overall 'flow of the game') making the 'number/order of actions in a game' equate to 'strategy in game.' Thinking that the two are somehow 'separate' is more of an illusion than anything else; a slower player will use a strategy that requires less actions than a faster player. Granted, we can have a scenario where a slower player is using a faster strategy with more actions than s/he can perform, but that sounds like a formula for possibly losing. And I've already assumed that the 'strategies' as defined are all 'to win.'
End Paragraph of Text
So far, we haven't added in speed to this, but let's take a base speed (we'll call this speed Normal). With normal, we have some number of strategies X consisting of specific series of actions at this speed. X in this case is finite; because humans can only make a max number of actions in a given span of time. So far so good, yeah? I'm pretty sure I haven't said anything really illogical at all. (Note: Tailoring X to 'all strategies including those that make you lose' does not fundamentally make it different from 'all strategies including those that allow you to win,' but I assume the latter case is what we're all most familiar with.)
Now, let's consider a case where the base speed is higher (we'll call this speed Fastest). We now need to consider how X has changed. Let's see. For X with many actions at Normal, some of them will be eliminated at higher speed because, unfortunately, they break the upper limit of human hand speed, yeah? So automatically, we get a limitation on X and X at fastest (hereby dubbed Y) has to be less than X. Already, we have a reduction in the empirical number of strategies.
However, this is not the only factor reducing X; SC, as many of us have pointed out ad nausem, is an e-Sport and a competitive game. Out of the new number Y (which, remember, is now the number of all possible strategies in time Fastest tailored to the upper limit of human physical capability) there has to be as least a few strategies (more than a few, actually) which are not viable for competitive play. I will define this selection process of viability as metagame selection. Because of SC's competitive nature, Y is reduced further to some smaller number. As you can see, raising the speed and considering SC as an e-Sport does decrease the empirical number of strategies.
The viewpoint most of the counter-arguments in this thread commonly misunderstand this. Strategic depth does decrease at higher speeds. This is a natural evolution of the game; it exists in anything that's a competitive environment when a given amount of possibilities (strategy for SC, game plan for MTG, course path for Mario Kart, move choice for Dead or Alive 4) correlates to a factor defined as an essential skill. However, to the writer of the editorial who probably isn't a 'competitive hardcore gamer,' so to speak, it is a bad thing; to us, possibly not. It does not change the fact that this is true.
Point 2) A lot of people in this thread have argued, extensively I might add, that somehow thinking 'on the fly' is an 'improper approach.' This, I have no problem with.
However, when this argument was blown into 'on the fly practically does not exist and all the strategy of a game *should be prepared in advanced*,' I have to question it. Just because someone aspires to be able to do things 'on the fly' is not a reason in and of itself to brand him simply as 'a low level gamer.'
As important as hard work is, and constantly practicing and honing our 'game sense' (read: ability to recognize patterns and respond to them with a set of pre-planned actions that we believe have the highest utility in that pattern), this is not actually 'game sense.' Sure, we are dazzled and awed by the amazing strategic and tactical skills that progamers have (hell, I love Flash), but how much of that is routine 'flow-charting' and how much of that is actually game sense? As you can probably tell by now, I am defining game sense much differently than the actual applied definition you are all used to; I am talking much more about the ability to accurately read and analyze a situation correctly (yes, even with factors taking in lack of information).
You are all saying that his lack of experience with the game contributed to his loss; you would all also be right. This is not what I'm contesting. What I am contesting is the fact that somehow, aspiring to aid the development in 'true game sense' (for Dan, it would be reducing game speed to make up for mechanical differences) makes someone a low-level player. It simply means that he wants a game where he can think more.
(Note* For those who are going to misinterpret this, I am not saying Dan isn't a low level player. I am merely saying that you cannot judge this simply on someone's opinion of a game. A dissention from the majoirty should not be the sole criteria for calling someone 'low leveled,' 'n00bish,' and whatever names those wonderful emails in his mailbox probably have. I have, however, no objection to calling him low level because he has no skill. Hell, neither do I.)
Point 3) The chess analogy is the worse analogy I've ever heard. A more comparable analogy is comparing Normal/Fastest SC to Normal/Fastest speeds in an FPS. Either way, Dan's point of mechanics mattering more than strategic depth (e.g. shooting someone in the head vs shooting someone period) remains valid. I'm pretty sure none of you can say it's wrong; you can say it is meaningless, but it isn't wrong.
Point 4) I guess I should finally get onto my opinions on the article itself.
Dan isn't wrong. His points are pretty clear, and generally, pretty logical (as I've stated my reasonings for believing so above). Strategic depth depreciates with game speed due to the creation of a 'metagame.' The emphasis of mechanics generally is greater than 'natural game sense' at higher speeds than lower speeds, tying the concepts of physical ability and strategy into one neat little box. Dan is not wrong about what he says.
Then what is he wrong about? I'm not really sure. A lot of you do take a superior tone, albeit it is justified with all of the hard practicing, thread reading, and VOD-watching that you've all been doing, but in the end, Dan is a casual player. His approach to the game fails because what he has described is fundamentally not an e-Sport. I don't think I need to go on with this line of thought; it's simply, as a previous poster said, a 'culture clash.' Does it make Dan wrong? No. Does it make us right? No. Just two different schools (and frankly, I'm very surprised why this topic had garnered as many responses as it did). Let him play games his way, and keep playing games your way. I must admit, this situation is paralleled very well in just about everything even mildly competitive, from WoW to like, poker.
Some of you have tried to list ideas on how to 'heal' this rift or somehow, not make it so much of a divisive issue. Unfortunately, I doubt that is really possible. Just judging from the number of people jumping down an editorial writer's throat to this extent because he said something we all disagree with (and forget about the MBS threads), the SC scene probably has a long ways to go. It's probably a good thing that the majority of SC players, as opposed to other things like WoW and MTG, are competitive, but it also makes it hard for the lone casual kid who likes building nukes because he thinks it's fun (and gets whipped all the time for it). (Note: I hate Terran because I suck with Terran. But Nukes are awesome.)
My personal view on how to 'fix' something like this is if SC was a 'game without a metagame.' A competitive game without a defined metagame. Anything and everything is viable. The fact that SC itself does not have a world vast enough to make this possible (not to mention the logistics for an RTS of this sort) makes this sort of fix pretty much ... impossible, but just for food for thought, here's an close example to what sort of system I might mean.
Imagine a competitive version of Spore. =D
-S3ra
PS: Flames will be met with cookies.
|
MyLostTemple
United States2921 Posts
i really hope he actually reads this thread.
|
On May 07 2008 10:16 teamsolid wrote:Show nested quote +On May 07 2008 09:50 Day[9] wrote: to anyone who has some high level undergraduate discrete mathematics courses in operations research:
excel can't exactly vomit out the optimal solution with a handful of numbers. you have to cleverly set up a linear program w/ the proper constraints. since linear programs can be solved in polynomial time, this is good news.
however, its extremely difficult to setup a linear program to solve an RTS problem. for example, resources are particularly hard to model. For example, the rate of mineral increase is a step function that behave in a linear fashion (when you have more mineral patches than workers) or in a logarithmically diminishing pattern (when you slowly have more SCVs than minerals) and have seemingly peculiar spikes/changes (when you transfer your workers to an expansion). As a result, even should you obtain an excellent model for resource collection, you still have a non-linear constraint on your hands, yielding a non-linear programming problem.
moreover, since RTS game are inherently discrete, you additionally have an integer programming problem. Occasionally you can circumvent this problem by letting your variables be time variables denoting the time at which unit/building X is constructed. Naturally, your objective function will be minimizing the time of the "last" built unit. However, you'll have a bit of an issue with "army" variability, as you'll need to specify exactly which units/buildings are going to be built (remember your variables are time variables).
I spent much of last semester constructing a linear programming problem to construct the optimal fast tank push build. My team accounted for the integer nature of the problem, and used the "time variables" i mentioned in the previous problem. That said, we had to assume infinite resources, as it was far too difficult (given our time span) to appropriately model a resource function. Even if we DID get a good model, we'd have a non-linear integer programming problem which is a real bitch to solve. Even with our simple model of "you have infinite resources, we need to build these specific units/buildings in the minimum amount of time," we still had alot of difficulty properly inputting our constraints.
I'd love to look at some of these "magical excel spreadsheets" that produce optimal values. But, considering the experience I have in modelling such things, I can't quite see how such a task would be feasible unless the player was a talented mathematician/engineer.
Thoughts? You do realize that Dan just made that whole excel-sheet story up right? Even using the term "Min/Maxer" is pretty stupid, because you might as well replace it with "good player" or "pro".
Meh, I don't think there's any need to note whether he made it up or not. I think he has an entirely valid concern that some could use out-of-game mathematical techniques to optimize particular situations, optimizations which could take away from the beauty/joy/strategy of the game. That said, I want to make it equally clear that optimizing in an eye-opening, useful way is a virtually intractable problem (as far as I can imagine. maybe there's some suave mathematician who's constructed a genius model that can be solved in polynomial time).
Either way, whether he made up the story or not, mathematical optimization techniques can't really be part of a sound argument against speed in games.
Hmmm... I feel an article coming on...
|
On May 07 2008 09:50 Day[9] wrote: to anyone who has some high level undergraduate discrete mathematics courses in operations research:
excel can't exactly vomit out the optimal solution with a handful of numbers. you have to cleverly set up a linear program w/ the proper constraints. since linear programs can be solved in polynomial time, this is good news.
however, its extremely difficult to setup a linear program to solve an RTS problem. for example, resources are particularly hard to model. For example, the rate of mineral increase is a step function that behave in a linear fashion (when you have more mineral patches than workers) or in a logarithmically diminishing pattern (when you slowly have more SCVs than minerals) and have seemingly peculiar spikes/changes (when you transfer your workers to an expansion). As a result, even should you obtain an excellent model for resource collection, you still have a non-linear constraint on your hands, yielding a non-linear programming problem.
moreover, since RTS game are inherently discrete, you additionally have an integer programming problem. Occasionally you can circumvent this problem by letting your variables be time variables denoting the time at which unit/building X is constructed. Naturally, your objective function will be minimizing the time of the "last" built unit. However, you'll have a bit of an issue with "army" variability, as you'll need to specify exactly which units/buildings are going to be built (remember your variables are time variables).
I spent much of last semester constructing a linear programming problem to construct the optimal fast tank push build. My team accounted for the integer nature of the problem, and used the "time variables" i mentioned in the previous problem. That said, we had to assume infinite resources, as it was far too difficult (given our time span) to appropriately model a resource function. Even if we DID get a good model, we'd have a non-linear integer programming problem which is a real bitch to solve. Even with our simple model of "you have infinite resources, we need to build these specific units/buildings in the minimum amount of time," we still had alot of difficulty properly inputting our constraints.
I'd love to look at some of these "magical excel spreadsheets" that produce optimal values. But, considering the experience I have in modelling such things, I can't quite see how such a task would be feasible unless the player was a talented mathematician/engineer.
Thoughts?
It doesn't have to be discrete, a resource function or almost anything else for that matter could be fairly well approximated by something continuous.
The optimal builds in starcraft weren't found by excel, they were found by trial-and-error. For example, building a pylon when 7 probes are built and the 8th is building. Regardless of how they were found though, his point still stands (somewhat): everyone has access to the best build orders just by looking up the "literature". But the this isn't what the strategy is about anymore.
|
On May 07 2008 12:56 s3raph wrote: First time poster, long time reader. Just wanted to respond to the thread with a few of my thoughts.
.....
-S3ra
PS: Flames will be met with cookies.
Excelent post chap I'm with you.
S.
|
+ Show Spoiler +On May 07 2008 12:56 s3raph wrote: First time poster, long time reader. Just wanted to respond to the thread with a few of my thoughts.
Point 1) First, on the issue of strategic depth vs speed. I'm pretty sure that I won't need to define the word 'strategy' to anyone within this thread, but what does this strategy actually consist of in-game? Please, something we can count; and that would be the actions performed. We can make the case that the actions a player takes within a game correspond with his strategy; furthermore, we can make the corollary that for each 'strategy' there is a specific series or pattern of actions that takes place. You can obviously expand this as you like, but for my purposes, all we need to remember is that there are some number of strategies that are correlated by a specific series of actions. Let's call the number of strategies X.
Begin Text for Clarification
Alright, so, a lot of you are defining 'strategy' as 'general overall flow of the game.' However, we can equate this to what actions we perform (we take with it the assumption that all actions taken in a game of StarCraft correspond to the overall 'flow of the game') making the 'number/order of actions in a game' equate to 'strategy in game.' Thinking that the two are somehow 'separate' is more of an illusion than anything else; a slower player will use a strategy that requires less actions than a faster player. Granted, we can have a scenario where a slower player is using a faster strategy with more actions than s/he can perform, but that sounds like a formula for possibly losing. And I've already assumed that the 'strategies' as defined are all 'to win.'
End Paragraph of Text
So far, we haven't added in speed to this, but let's take a base speed (we'll call this speed Normal). With normal, we have some number of strategies X consisting of specific series of actions at this speed. X in this case is finite; because humans can only make a max number of actions in a given span of time. So far so good, yeah? I'm pretty sure I haven't said anything really illogical at all. (Note: Tailoring X to 'all strategies including those that make you lose' does not fundamentally make it different from 'all strategies including those that allow you to win,' but I assume the latter case is what we're all most familiar with.)
Now, let's consider a case where the base speed is higher (we'll call this speed Fastest). We now need to consider how X has changed. Let's see. For X with many actions at Normal, some of them will be eliminated at higher speed because, unfortunately, they break the upper limit of human hand speed, yeah? So automatically, we get a limitation on X and X at fastest (hereby dubbed Y) has to be less than X. Already, we have a reduction in the empirical number of strategies.
However, this is not the only factor reducing X; SC, as many of us have pointed out ad nausem, is an e-Sport and a competitive game. Out of the new number Y (which, remember, is now the number of all possible strategies in time Fastest tailored to the upper limit of human physical capability) there has to be as least a few strategies (more than a few, actually) which are not viable for competitive play. I will define this selection process of viability as metagame selection. Because of SC's competitive nature, Y is reduced further to some smaller number. As you can see, raising the speed and considering SC as an e-Sport does decrease the empirical number of strategies.
The viewpoint most of the counter-arguments in this thread commonly misunderstand this. Strategic depth does decrease at higher speeds. This is a natural evolution of the game; it exists in anything that's a competitive environment when a given amount of possibilities (strategy for SC, game plan for MTG, course path for Mario Kart, move choice for Dead or Alive 4) correlates to a factor defined as an essential skill. However, to the writer of the editorial who probably isn't a 'competitive hardcore gamer,' so to speak, it is a bad thing; to us, possibly not. It does not change the fact that this is true.
Point 2) A lot of people in this thread have argued, extensively I might add, that somehow thinking 'on the fly' is an 'improper approach.' This, I have no problem with.
However, when this argument was blown into 'on the fly practically does not exist and all the strategy of a game *should be prepared in advanced*,' I have to question it. Just because someone aspires to be able to do things 'on the fly' is not a reason in and of itself to brand him simply as 'a low level gamer.'
As important as hard work is, and constantly practicing and honing our 'game sense' (read: ability to recognize patterns and respond to them with a set of pre-planned actions that we believe have the highest utility in that pattern), this is not actually 'game sense.' Sure, we are dazzled and awed by the amazing strategic and tactical skills that progamers have (hell, I love Flash), but how much of that is routine 'flow-charting' and how much of that is actually game sense? As you can probably tell by now, I am defining game sense much differently than the actual applied definition you are all used to; I am talking much more about the ability to accurately read and analyze a situation correctly (yes, even with factors taking in lack of information).
You are all saying that his lack of experience with the game contributed to his loss; you would all also be right. This is not what I'm contesting. What I am contesting is the fact that somehow, aspiring to aid the development in 'true game sense' (for Dan, it would be reducing game speed to make up for mechanical differences) makes someone a low-level player. It simply means that he wants a game where he can think more.
(Note* For those who are going to misinterpret this, I am not saying Dan isn't a low level player. I am merely saying that you cannot judge this simply on someone's opinion of a game. A dissention from the majoirty should not be the sole criteria for calling someone 'low leveled,' 'n00bish,' and whatever names those wonderful emails in his mailbox probably have. I have, however, no objection to calling him low level because he has no skill. Hell, neither do I.)
Point 3) The chess analogy is the worse analogy I've ever heard. A more comparable analogy is comparing Normal/Fastest SC to Normal/Fastest speeds in an FPS. Either way, Dan's point of mechanics mattering more than strategic depth (e.g. shooting someone in the head vs shooting someone period) remains valid. I'm pretty sure none of you can say it's wrong; you can say it is meaningless, but it isn't wrong.
Point 4) I guess I should finally get onto my opinions on the article itself.
Dan isn't wrong. His points are pretty clear, and generally, pretty logical (as I've stated my reasonings for believing so above). Strategic depth depreciates with game speed due to the creation of a 'metagame.' The emphasis of mechanics generally is greater than 'natural game sense' at higher speeds than lower speeds, tying the concepts of physical ability and strategy into one neat little box. Dan is not wrong about what he says.
Then what is he wrong about? I'm not really sure. A lot of you do take a superior tone, albeit it is justified with all of the hard practicing, thread reading, and VOD-watching that you've all been doing, but in the end, Dan is a casual player. His approach to the game fails because what he has described is fundamentally not an e-Sport. I don't think I need to go on with this line of thought; it's simply, as a previous poster said, a 'culture clash.' Does it make Dan wrong? No. Does it make us right? No. Just two different schools (and frankly, I'm very surprised why this topic had garnered as many responses as it did). Let him play games his way, and keep playing games your way. I must admit, this situation is paralleled very well in just about everything even mildly competitive, from WoW to like, poker.
Some of you have tried to list ideas on how to 'heal' this rift or somehow, not make it so much of a divisive issue. Unfortunately, I doubt that is really possible. Just judging from the number of people jumping down an editorial writer's throat to this extent because he said something we all disagree with (and forget about the MBS threads), the SC scene probably has a long ways to go. It's probably a good thing that the majority of SC players, as opposed to other things like WoW and MTG, are competitive, but it also makes it hard for the lone casual kid who likes building nukes because he thinks it's fun (and gets whipped all the time for it). (Note: I hate Terran because I suck with Terran. But Nukes are awesome.)
My personal view on how to 'fix' something like this is if SC was a 'game without a metagame.' A competitive game without a defined metagame. Anything and everything is viable. The fact that SC itself does not have a world vast enough to make this possible (not to mention the logistics for an RTS of this sort) makes this sort of fix pretty much ... impossible, but just for food for thought, here's an close example to what sort of system I might mean.
Imagine a competitive version of Spore. =D
-S3ra
PS: Flames will be met with cookies.
"The viewpoint most of the counter-arguments in this thread commonly misunderstand this. Strategic depth does decrease at higher speeds."
No, this is not true. Higher speed decreases the ammount of strategies that rely on good mechanics, but at the same time it increases the number of strategies that rely on that fact that one's opponent can't pay attention to everything (e.g. drops, hold-Lurkers, etc.). So the higher speed only alter the balance between the two types of strategy.
Also higher speed means that mechanics get harder, which in turn makes them less important in relation to strategy.
What's more important, is the fact that higher speed introduces a whole new layer of strategical thinking - time & attention management. This simply does not exist when playing at slower game pace.
To sum up, you're basing this point on a flawed assumption that higher speed reduces a number of one type of strategies but doesn't increase the amount of another type of strategies, which is actually the case. You've said you are a long tiem reader, but apparently you weren't reading closely, as it (the fact that higher speed introduces time/attention management and introduces a new type of strategies) has already been pointed out.
"A lot of people in this thread have argued, extensively I might add, that somehow thinking 'on the fly' is an 'improper approach.' This, I have no problem with.
However, when this argument was blown into 'on the fly practically does not exist and all the strategy of a game *should be prepared in advanced*,' I have to question it. Just because someone aspires to be able to do things 'on the fly' is not a reason in and of itself to brand him simply as 'a low level gamer.'"
The fact is that pre-planned strategies >>> improvisation. It's very easy spot holes in the strategy of the player improvising, whereas it's close to impossible to spot them in a strategy that has been perfected for months if not years. That is the reason why players who don't recognize that fact are labled as 'low level gamers'.
"It simply means that he wants a game where he can think more."
He wants to have more time to improvise, i.e. do something that is counter-productive...
"A more comparable analogy is comparing Normal/Fastest SC to Normal/Fastest speeds in an FPS. Either way, Dan's point of mechanics mattering more than strategic depth (e.g. shooting someone in the head vs shooting someone period) remains valid. I'm pretty sure none of you can say it's wrong; you can say it is meaningless, but it isn't wrong."
No, that's simply wrong. Mechanics outweigh strategy only when theere is a large gap in mechanics between the two players, and good mechanics usually go in pair with the understanding of strategy in SC. As we go higher in skill level, the mechanics begin to matter less and less. On progaming level it's all about strategy, tactics, and other types of decision making.
This renders Dan's point wrong as he was generalizing and therefore claiming his point hold true for all skill levels, which is simply not true.
|
On May 07 2008 21:34 Wonders wrote:Show nested quote +On May 07 2008 09:50 Day[9] wrote: to anyone who has some high level undergraduate discrete mathematics courses in operations research:
excel can't exactly vomit out the optimal solution with a handful of numbers. you have to cleverly set up a linear program w/ the proper constraints. since linear programs can be solved in polynomial time, this is good news.
however, its extremely difficult to setup a linear program to solve an RTS problem. for example, resources are particularly hard to model. For example, the rate of mineral increase is a step function that behave in a linear fashion (when you have more mineral patches than workers) or in a logarithmically diminishing pattern (when you slowly have more SCVs than minerals) and have seemingly peculiar spikes/changes (when you transfer your workers to an expansion). As a result, even should you obtain an excellent model for resource collection, you still have a non-linear constraint on your hands, yielding a non-linear programming problem.
moreover, since RTS game are inherently discrete, you additionally have an integer programming problem. Occasionally you can circumvent this problem by letting your variables be time variables denoting the time at which unit/building X is constructed. Naturally, your objective function will be minimizing the time of the "last" built unit. However, you'll have a bit of an issue with "army" variability, as you'll need to specify exactly which units/buildings are going to be built (remember your variables are time variables).
I spent much of last semester constructing a linear programming problem to construct the optimal fast tank push build. My team accounted for the integer nature of the problem, and used the "time variables" i mentioned in the previous problem. That said, we had to assume infinite resources, as it was far too difficult (given our time span) to appropriately model a resource function. Even if we DID get a good model, we'd have a non-linear integer programming problem which is a real bitch to solve. Even with our simple model of "you have infinite resources, we need to build these specific units/buildings in the minimum amount of time," we still had alot of difficulty properly inputting our constraints.
I'd love to look at some of these "magical excel spreadsheets" that produce optimal values. But, considering the experience I have in modelling such things, I can't quite see how such a task would be feasible unless the player was a talented mathematician/engineer.
Thoughts? It doesn't have to be discrete, a resource function or almost anything else for that matter could be fairly well approximated by something continuous.
I think you misunderstood what I said. A programming problem is discrete if the control variables are discrete in nature. Since our control variables are units/buildings, they are inherently discrete. (you can't say, build .698 of a marine or 1/2 a building or something, you have to choose exactly whether or not to build it). It is THIS discreteness that give you an integer programming problem.
As I said (perhaps not so clearly), a well modeled continuous resource function will be undeniably nonlinear (and oftentimes piece-wise defined), which will result in a non-linear constraint, yielding a non-linear programming problem.
Integer programming problems and non-linear programming problems are already hard to solve. Since I can't imagine an extremely flexible well done model that isn't non-linear and discrete in nature, I don't see any mathematical model optimizing our game any time soon
The optimal builds in starcraft weren't found by excel, they were found by trial-and-error. For example, building a pylon when 7 probes are built and the 8th is building. Regardless of how they were found though, his point still stands (somewhat): everyone has access to the best build orders just by looking up the "literature". But the this isn't what the strategy is about anymore.
My apologies if you misunderstood. I wasn't by any means saying that optimal builds are found in excel . I was merely addressing the Editorial's statement that "min/maxers use excel to optimize builds."
|
s3raph: First of all, I think your understanding of the game is probably about 10 times more thorough than Dan's. In fact, if you turned the post you made here into a PCGamer article, I don't think there would be any real backlash from SC players at all. I think you are arguing similar points to Dan, but at a much higher level and with reasonable arguments and examples. Now this is what Dan should've wrote if he simply wanted to "stimulate" discussion as he claims.
On May 07 2008 12:56 s3raph wrote: First time poster, long time reader. Just wanted to respond to the thread with a few of my thoughts.
Point 1) First, on the issue of strategic depth vs speed. I'm pretty sure that I won't need to define the word 'strategy' to anyone within this thread, but what does this strategy actually consist of in-game? Please, something we can count; and that would be the actions performed. We can make the case that the actions a player takes within a game correspond with his strategy; furthermore, we can make the corollary that for each 'strategy' there is a specific series or pattern of actions that takes place. You can obviously expand this as you like, but for my purposes, all we need to remember is that there are some number of strategies that are correlated by a specific series of actions. Let's call the number of strategies X. Okay, I think I understand what you're saying here. Basically, in order to be able to pull off many of the more advanced strategies in SC, you must achieve a minimum level of mechanics before they can become viable.
Examples: Sair/Reaver, Bisu build (effectively), SK Terran, Savior/JD-esque muta harass to hive tech
These all require a high degree amount of multitasking to pull off, before they can become more effective than a comparative standard/easy macro build. Perhaps these advanced builds would indeed become more viable for even lesser players to perform if the speed were lowered to Normal.
However, I think this really only applies to a low-level player who also has full understanding of strategies in the game. This is not true for 95% of players out there. Most of the time, as a player's understanding of the game improves, so do his mechanics along with them. The exception would be someone who spectates many E-Sport matches, but doesn't actually play much SC (perhaps more prevalent in Korea).
I guess the solution would just be for these players to train more and improve their mechanics by facing players of similar skill. It gives you a goal to reach at least and let's you marvel at how well the top gamers can play.
But again, I still don't think the actual strategical depth of the game itself is lowered at all, but it's just that the mechanics are proportionally more important as speed is increased, so this requires one to master both aspects of the game (mechanics AND strategy). Casual players on the other hand can just ignore all of this and play however they wish.
Finally, as mentioned by several posters, reducing the speed would also greatly reduce the importance of the 3rd resource in the game: time. Perhaps this can be considered more to be "tactics" than "strategy", but either way that reduces a significant aspect of the game.
On May 07 2008 12:56 s3raph wrote: You are all saying that his lack of experience with the game contributed to his loss; you would all also be right. This is not what I'm contesting. What I am contesting is the fact that somehow, aspiring to aid the development in 'true game sense' (for Dan, it would be reducing game speed to make up for mechanical differences) makes someone a low-level player. It simply means that he wants a game where he can think more. So what you're saying is basically that Starcraft would have a shorter learning curve (both strategically and mechanically) if it were played on Normal speed rather than fastest. You know, I do have to kind of agree with you there. But then again, I would argue that as long as both players are fairly evenly matched, I don't think there is any extra pressure to focus solely on mechanics at the expense of "game sense" and "strategy".
I think this problem only exists with the outdated matchmaking of SC's Battle.net where there's often a large skill gap between two players, and it could be entirely eliminated with a proper AMM. It is however very intimidating for a new player to jump into a 10-yr old game like SC where the average skill level is now ridiculously high. Many casual gamers would simply avoid 1v1 because they claim it's "too intense". This should not, however, be a problem in a freshly released game like SC2 and a fully functional AMM.
On May 07 2008 12:56 s3raph wrote: Point 3) The chess analogy is the worse analogy I've ever heard. A more comparable analogy is comparing Normal/Fastest SC to Normal/Fastest speeds in an FPS. Either way, Dan's point of mechanics mattering more than strategic depth (e.g. shooting someone in the head vs shooting someone period) remains valid. I'm pretty sure none of you can say it's wrong; you can say it is meaningless, but it isn't wrong. Dan must've worded his article pretty awfully then, because I got a whole different message from that. I'm too lazy to find quotes right now, but I do think that he is actually was wrong in many regards.
On May 07 2008 12:56 s3raph wrote: Point 4) I guess I should finally get onto my opinions on the article itself.
Dan isn't wrong. His points are pretty clear, and generally, pretty logical (as I've stated my reasonings for believing so above). Strategic depth depreciates with game speed due to the creation of a 'metagame.' The emphasis of mechanics generally is greater than 'natural game sense' at higher speeds than lower speeds, tying the concepts of physical ability and strategy into one neat little box. Dan is not wrong about what he says. This part I would have to disagree with. I would agree that mechanics are more highly emphasized (i.e. require more training) at fastest speed than at a lower speed, but that does not reduce the total amount of strategy in the game. This is evident when you compare a foreigner with 300 APM to a player like Savior (~250 apm). In fact "game sense" and "strategy" plays a huge role in skill level. Myself, I average around ~120-130 apm in games but I regularly defeat players with 200+ APM on ICCUP all the time (I'm at C- level).
As for your "nuke" example and the casual gamer. So long as the casual gamer is playing against another of about equal skill or understanding of the game, he can have all the fun he wants with nukes, sci vessels, BCs, valks, wraiths etc. However, even if the game speed were reduced to normal, I still doubt that the nuke would suddenly become viable at higher levels, simply because it can be easily countered and it requires the replacement of a comsat. This is more due to the balance in the game rather than its speed.
I guess his article is correct if he talks exclusively about players who are entirely new to RTS's in general, the players who have apm's in the range from 20-60 (I'm not sure how long it takes people to advance beyond this level). However, this is NOT what he did. Nowhere did he refer or mention that his article speaks only about the lowest-skilled players, and in fact he even applies this to pro-gamers.
His mistake is that he made sweeping, generalized conclusions about the game as a whole and its players based on his superficial view of the game. As a basic rule of thumb, you should never make judgments about any midly competitive sport or game by looking at how it's played at the lowest level of skill. Instead, the merit of game should be based on how it's played by the very best.
Using the example that was previously mentioned, a low level Chess player (like me) simply sees the game as trading pieces until one player ends up on top by gaining an extra piece. Do I believe that Chess is devoid of strategy because I do not understand it? No, I realize that it is my own judgment and understanding of the game that is lacking.
|
I'm surprised there isn't some korean program that mathematically works out a few of the simple build orders given that there's a lot of money in making perfect builds i would have thought the progamers would be interested in this. And given the inherently computer savvy fans you'd have thought there'd be something by now, a comp sci masters thesis turned pro team tool.
|
|
|
|